In general, the way an object is reflected depends on the light source. This is due to the color temperature that is different for each light source. For example, a white object appears reddish when placed under a light source of low color temperature. On the other hand, when the object is placed under a light source of high color temperature, the object appears bluish.
For this reason, most photographing apparatuses, such as a camera, compensate the difference of color, caused by different color temperatures, by use of a method called automatic white balance (AWB).
Below, the automatic white balance method in accordance with the prior art is briefly described.
First, the picture data generated through an image sensor in a photographing apparatus is transformed to color signals (R, G, B), and then R and B gains are adjusted. Then, the adjusted color signals are transformed to the color space of brightness (Y) and color difference signals (R-Y, B-Y), and the color difference signals of a screen are integrated and the average of the integration is calculated. This average is assumed to be the white value.
If the derived average value changes from the assumed white value due to the change of external light source, the changed R and B gains are calculated. By adjusting the R and B gains on the changed R and B gains, the white balance is adjusted.
FIGS. 1-3 illustrate conditions for detecting color temperature for applying the automatic white balance method of the prior art.
As shown in each of FIGS. 1-3, a color zone (the filtering zone for detecting valid color information only) is set such that a characteristic curve 110 is included on a UV plane. The characteristic curve 110 is a curve calculated by coordinates, in which color information (R, G, B) from reflecting a color on a virtual object of achromatic color (e.g. white, gray, and black) is present. The red component becomes intense near the upper end 115 of the characteristic curve, and the blue component becomes intense near the lower end 120. That is, the image data has a reflector, which is achromatic and thus makes the image data have the same color as the light source, and, by determining this color, the current color temperatures can be inferred. YUV is a way of expressing the color, breaking down to Y component, which is luminance, and U and V components, which are chrominance. If R, G, and B values are given, Y, U, and V can be obtained from the following Eqs. 1-3:Y=0.3R+0.59G+0.11B  Eq. 1U=(B−Y)×0.493  Eq. 2V=(R−Y)×0.877  Eq. 3
The color zone is an area on the UV plane that has high possibility of reflected light by the reflector. Any light outside the color zone can be considered to be not of reflected light by the reflector. Therefore, the most important factor for determining color temperature can be determining the area and shape of the color zone.
In the prior art, the color zone for the determination of color temperature has been determined to be a rectangle 130 containing all of the characteristic curve 110 (see FIG. 1), two rectangles 210 and 220 (i.e. one rectangle 210 containing the upper characteristic curve and the other rectangle 220 containing the lower characteristic curve, about the origin), or a parallelogram 310 containing all of the characteristic curve 110. Then, the color temperature in a frame has been determined and corrected, using color information corresponding to the pertinent color zone.
However, when the color zone is formed by one rectangle 130, as shown in FIG. 4, color information 410 and 420 that is unrelated to the characteristic curve is included, causing errors in detecting color temperature. Although a plurality of rectangles 210 and 220 or a parallelogram 310 can be used to form the color zone, in order to overcome this problem, discontinuity of color zones or unnecessary color information cannot still be avoided.
Moreover, considering that the characteristic curve of the image sensor of a photographing apparatus is a little more distorted than normal characteristics, the various conventional methods for determining the color zone still have a problem of containing more erroneous color information (i.e. color information, included by unnecessarily expanding the color zone, which increases error rates while detecting color temperatures). This increases the error rates when detecting color temperatures and increases the amount of computation, making the color temperature detecting apparatus more complicated.
The conventional method of detecting color temperature interpolates the RGB Bayer type image to derive a RGB value, which is then converted to a YUV value, and the color information using this YUV value to match on the UV plane is used. However, detecting color temperature using the YUV value had problems of the structure being too complicated and the calculation increasing, due to the calculations of decimal numbers, negative numbers, and multiplication.